Please use this identifier to cite or link to this item:
Authors: Sayanti Guha Majumdar
Advisor: Anil Rai
Title: Development of Integrated Model forGenomic Selection
Language: en_US
Type: Thesis
Agrotags: null
Keywords: Genomic Selection; Linear Model; Non-linear Model; Error Variance; Imputation
Abstract: Genomic Selection (GS) is a recent area for efficient breeding of animals and plants. GS has been used globally for increasing agricultural production and productivity in recent days. It is suitable for selecting complex quantitative traits which lead to the efficient selection of breeding material after predicting Genomic Estimated Breeding Values (GEBVs) of target species. The accuracy of estimation of GEBVs depends on a large number of factors which include training population, genetic architecture of target species, statistical models, etc. Accuracy of selection of breeding parents also varies based on selected GS models according to their assumptions and treatments of marker effects. The first step of GS is feature (marker) selection. There are several models available in the literature for selection features in GS. However, applicability of these models is based on many factors including extent of additive and epistatic effects of breeding population. Therefore, there is strong need to evaluate the performance of these models and techniques of feature selection under different genetic architecture. Furthermore, statistical models for Genomic Selection available for estimation of Genomic Estimated Breeding Value are not robust against this genetic architecture and depend on datasets. Some models perform well for additive genetic architecture and others perform well for non-additive genetic architecture. But, there is lack of estimator which could capture both of these effects simultaneously. Therefore, this study has been conducted to develop a robust estimator which may be able to capture both additive and non-additive effects efficiently. First, the performance of linear/ additive effect models as well as nonlinear/epistatic effect models have been evaluated through a simulation study. In general, performance of SpAM was found to be superior for GS than all other additive effects models considered in this study. However, in case of low heritability and high epistatic effect, the HSIC LASSO out performed all competitive models. Therefore, an robust integrated estimator has been developed by combining these two efficient additive and non-additive models i.e. SpAM and HSIC LASSO respectively. Further, for estimation of error variance four different methods have been evaluated which are being used for estimation of weight in the developed Integrated Model. The performance of the proposed model has been evaluated on the basis of prediction accuracy, fraction of correctly selected features and redundancy rate along with their standard error of mean. Further, the performance of the proposed model has been Abstract 78 compared with SpAM and HSIC LASSO with respect to the above criteria. The newly developed estimator is found to be superior in terms of its performance and it has been demonstrated to be robust against any genetic architecture of datasets. Also the performance of the developed Integrated Model has been evaluated in case of 2%, 5% and 10% genotypic imputation of data and it is found to be comparable with respect to the complete dataset. Keywords:
Description: T-10083
Subject: Bioinformatics
Theme: Development of Integrated Model forGenomic Selection
These Type: Ph.D
Issue Date: 2019
Appears in Collections:Theses

Files in This Item:
File Description SizeFormat 
full thesis_phd_sayanti.pdf
  Until 2020-09-11
1.62 MBAdobe PDFView/Open Request a copy

Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.